Minienvironments vs. ballrooms; challenging ULPA filters and more
Q: If minienvironments are so much better than Class 1 ballrooms, why isn`t everybody using them?
M. Southworth, Worcester, MA
A:You`ve really picked a controversial issue! It should be easier to keep a small, “mini” enclosure cleaner than a big production floor filled with people and machinery. There have been lots of experiments, documented thoroughly, with legitimate data showing that mini-environ ments work. Few people dispute the general conclusions of these experiences, and because of that, most of our clients have at least seriously considered minienvironments, if they have not converted either a little or lot of their production to them.
However, you asked “why isn`t everybody doing it?” There are several contributing factors. First, while some of the earlier recent applications (I say recent because long before full-coverage HEPA-filtered rooms were used, glove boxes achieved cleanliness levels much greater than the ambient environment) had high success rates of lowering contamination levels and increasing yields, they also came with built-in handicaps. Accessibility to the workstations using minienvironments was initially, and still is, an issue.
Unfortunately, I believe we tried to automate wafer handling inside of minienvironments faster than we were able to provide the necessary reliability of the automation equipment. This made accessibility an even bigger issue. The interface problems between the tool and the wafer enclosure were horrible and caused wafer breakage, leading to compromises about opening the enclosures to provide maintenance service in an environment which was much dirtier than we would have had in either a laminar flow cleanroom or inside the minienvironment. With good integration and much more reliable automation, most of these problems are behind us now.
Another disappointment has been the unrealized expectation that the initial cost of mechanical systems would be much lower, saving some big bucks. There are some suggestions that this should be a driver, but those who have really done it haven`t found it to be so. While it`s true that the cost for recirculation systems and full-ceiling filters is lower, the issue of keeping the wafers in a clean pod and transferring them into process tools has eaten up most of those savings. Furthermore, many of the manufacturers who have gone to minienvironments have wanted to do so with a capability of reverting back to a full-room cleanroom if it doesn`t work, and the initial costs have actually been higher than a conventional cleanroom. There`s no reason to think that will always be the case, but it has contributed to the reluctance for change.
Perhaps a bigger reason that minienvironment fabs haven`t swept through the industry faster is because many of the more successful manufacturers have been profitable enough that they can continue to build the big ballroom fabs and pay the higher operating costs to use them. Changing to minienvironment production is still a potential paradigm shift which has not become a widespread economic necessity. They are more suitable for older facilities where they provide a means to achieve substantially improved contamination control without building a new fab building, and they have been used extensively in this application.
Where minienvironments have made a major impact on the industry is in their direct addition to processing tools. Most of the contamination on the wafers doesn`t come from the cleanrooms — it comes from the processes and the process tools. Tool manufacturers know that, and while minienvironments are not a panacea for solving all of these problems, relative to the overall cost of the tools, they are an inexpensive approach to reduce the tool-induced contamination. So when you build your wafer fab, either by conventional or minienvironment design, and then spend the really big bucks on the tools, you get a lot of minienvironments anyway.
Q: Why do I need to challenge my ULPA filters to prove they are good? Doesn`t this reduce the efficiency of the filters and add contamination in the cleanroom?
R. F. Spencer, Champaign, IL
A: You have to put millions of particles above the filter in order to find and repair small leaks in properly manufactured ULPA filters. Otherwise, the time and effort would be much greater before starting to use a new cleanroom. If you don`t repair these leaks, the integrity of the cleanroom is jeopardized. While a properly defined challenge very slightly increases the pressure drop of the filters, it doesn`t appreciably shorten the expected lifetime of the filter, and it actually increases its efficiency. However, care must be given to select the proper type of challenge material; the material must not affect any downstream processing when the material comes through the leaks and through those which may occur later.
Q: How much cleaner will my fab be if I use 99.99999 percent ULPA filters?
S. Gibson, Portland, OR
A:My response to this question is bound to cause some controversy, but here goes: using “seven nines” or 99.99999 percent ULPA filters provides very little, if any, assurance that your cleanroom will be cleaner than it would be if you used properly manufactured, installed, and tested 99.99 percent or 99.999 percent HEPA filters. Contrary to what the performance efficiency numbers would tend to imply, you can get some horrible filters which pass such a standard. That`s because, in my opinion, the filter efficiency testing procedure, traditionally defined by UL586, is completely outdated for modern laminar flow cleanrooms. It is really an average testing procedure which can easily permit a large hole in an otherwise very good filter, allowing a stream of contamination to flow into your cleanroom where you don`t want it.
This test procedure is performed by equipment too cumbersome to be brought into an operating cleanroom, and is generally applied by the manufacturer to the filter media only, and then only on a sample basis. For good filters, most of the leaks don`t occur in the media. They are found most commonly at the edge of the filters due to sealing problems in manufacture or during filter installation. If you want assurance of a clean fab, pick a reputable manufacturer, ensure that the filters are scanned at the factory to a lower penetration level than you require in your cleanroom, impose a sample test upon receipt of the filters, and insist on a thorough filter-in-place leakage test after installation, using an appropriate challenge material, by a qualified third-party certifier.
Q: My company requires me to take a drink of water before I enter a cleanroom. Why?
J. Whitney, San Jose, CA
A:Didn`t your mother ever warn you to stop kissing your girlfriend or boyfriend because the interior of your mouth was a hotbed of germs and contamination? Well, it`s true. The inside of your mouth is filthy. This is especially true for smokers, who continue to exhale contamination for as long as twenty minutes after finishing that last cigarette. One way to reduce the contamination is to take a drink before going into the cleanroom. That way, all of that filth goes down into your stomach instead of into the cleanroom.
Q:Why do I have to have to follow a clean construction protocol? Can`t I save a lot of money by just using common sense construction? Has anyone ever done a real cost/benefit analysis of this fad?
E. Burnes, Chaska, MN
A: This is my pet peeve. You don`t have to use clean construction protocol (CCP). Many cleanrooms have been built in a conventional fashion, then cleaned up to perform their functions satisfactorily. Furthermore, any protocol which doesn`t use common sense isn`t a good one. Don`t misunderstand me, I believe in building clean. With the enormous investment firms are making in state-of-the-art wafer fabs, a well-managed protocol is a small price to pay to reduce the risk that the cleanroom won`t start up on schedule. It also reduces the risk that there will be some insidious sources of contamination that will have just a small impact on yields or on the ability to use the cleanroom for a more demanding process in the future. Yes, there have been a number of legitimate cost/benefit analyses on this practice, and I reported on some of them in a paper presented at both CleanRooms West `95 and CleanRooms East `96. However, these analyses generally do an objective job of determining the cost, while the ability to assign a “benefit” value is rather subjective. For more discussion on how to determine what protocol is appropriate for your construction, see Cleanrooms, the Protocol, an excellent book by Jerry Greiner.
Q:We are having an argument about what material is needed for ductwork for Class 1 air. I say stainless steel is needed; my boss says that if the big guys use galvanized, it should be good enough for us. Who is right?
G. Rosenstein, Scottsdale, AZ
A: Maybe the big guys are not right. (But remember, the boss is still the boss.) Class 1 air can be easily achieved with galvanized ductwork. It does make sense to use stainless immediately downstream of air handlers where moisture is injected to control humidity. However, there is a case for using stainless steel for the whole recirculation system. Ask your boss if he has ever been inside a return air duct servicing a cleanroom where there are acid clean stations, especially after it has been in operation for about a year. It is not unusual to have them look like the roof of the farmer`s barn that`s been out in the elements for 30 years. While we try to segregate the corrosive exhaust, some of the fumes always get sucked up into the clean air return. The “rust” inside the ductwork is loaded with iron and zinc particles, elements which are major lifetime killers in semiconductors. Certainly the filters will stop most of it. But filter systems aren`t perfect. In my experience in working with both stainless and galvanized systems, the choice is a function of the risks our clients are willing to take. Powder-coated epoxy paint on the interior of the ductwork is a reasonably-priced alternative.
Questions submitted for this month`s Ask the Experts column were answered by Donald L. Wadkins of Jacobs Engineering
Group. Wadkins is a leading industry expert on cleanroom protocol programs, with over 27 years of experience in semiconductor processing, cleanroom protocol management, and high-purity/process related systems. He is a member of the Technology Group at Jacobs Engineering Group`s Regional Office in Portland, OR, where he provides technical leadership and consultation to both internal and external clients. Wadkins has published a myriad of technical papers on cleanroom protocol, and regularly lectures at universities, corporations and other professional organizations on the subject of microcontamination control and protocol management.